[pkg-lua-devel] Bug#1101238: lua-lgi: FTBFS: tests/pango.lua:43: failure

[Chris Hofstaedtler]

On Tue, Jul 15, 2025 at 11:33:30PM +0200, Chris Hofstaedtler wrote:
> Hi Reiner,
> 
> On Tue, Jul 15, 2025 at 11:10:25PM +0200, Reiner Herrmann wrote:
> > Hi Chris,
> > 
> > On Tue, Jul 15, 2025 at 11:04:00PM +0200, Chris Hofstaedtler wrote:
> > > On Tue, Mar 25, 2025 at 12:11:09AM +0000, Santiago Vila wrote:
> > > > Package: src:lua-lgi
> > > 
> > thanks for bisecting.
> > But it seems to be inaccurate. We already found out (in the linked upstream
> > bug tracker) that pango 1.56.2 causes the regression.
> 
> Yeah, you're right. My earlier bisect failed due to unrelated 
> reasons. The real result is:
> 
> bisection finished successfully
>   last good timestamp: 20250313T150353Z
>   first bad timestamp: 20250313T210003Z
> the following packages differ between the last good and first bad timestamp:
>   gir1.2-pango-1.0:arm64 1.56.1-1 -> 1.56.2-1
>   libcairo-gobject2:arm64 1.18.2-2 -> 1.18.4-1
>   libcairo-script-interpreter2:arm64 1.18.2-2 -> 1.18.4-1
>   libcairo2-dev:arm64 1.18.2-2 -> 1.18.4-1
>   libcairo2:arm64 1.18.2-2 -> 1.18.4-1
>   libcap2:arm64 1:2.75-1 -> 1:2.75-2
>   libpango-1.0-0:arm64 1.56.1-1 -> 1.56.2-1

https://gitlab.gnome.org/GNOME/pango/-/merge_requests/829
looks like the "smoking gun". Especially the change in fwupd looks 
relevant: https://github.com/fwupd/fwupd/issues/8587
https://github.com/fwupd/fwupd/commit/d0d3eff666d2a6d9c09346ea050eed1c78271f0a

Chris